ADLC Phase Validation

VerifiedCaution

Validate any ADLC phase output before advancing to the next phase. Ensures quality and consistency across requirements, architecture, tasks, and implementation.

Sby Skills Guide Bot
DevelopmentIntermediate
7/28/2026
Claude Code
#adlc#validation#quality-gate#requirements#workflow

Recommended for

Automated summary

Checks the quality of ADLC phase artifacts (spec, architecture, tasks, implementation) before letting work advance to the next phase.

Strengths

  • Ships separate, detailed checklists for each of the four ADLC phases
  • Auto-detects the current phase from the artifacts present under .adlc/specs/REQ-xxx-*/
  • Covers cross-cutting consistency: acceptance-criteria coverage, cycle-free dependency DAGs, adherence to existing conventions and ADRs
  • Supports multi-repo setups by validating each task's repo: field against .adlc/config.yml

Limitations

  • Tightly bound to the .adlc/ layout and to a specific stack (Firestore, routes → services → repositories layering)
  • Validation is a model-read checklist: it does not actually execute the test suite or a linter
  • The provided SKILL.md is truncated at the reporting step, so the exact output format is not fully documented
When to use it

Use it right after drafting or updating an ADLC artifact, when you need a go/no-go decision before moving to the next phase.

When not to use it

Skip it on projects that don't follow the .adlc/ structure, since the skill stops immediately when .adlc/specs/ is missing.

Security analysis

Caution
Quality score88/100

The skill performs legitimate validation but uses shell command substitution to source an external script and inspect files. The repo-controlled ethos script introduces a supply-chain risk: if the ADLC repository is compromised, arbitrary commands could execute when the skill runs. No destructive or exfiltration actions are directly instructed, so it is not 'risky', but the use of shell execution for context loading warrants caution.

Findings
  • •Executes shell script from repo-controlled path `.adlc/partials/ethos-include.sh` via command substitution, which could run arbitrary code if the repository is malicious.
  • •Uses shell command substitution (`!` syntax) without declared allowed-tools, potentially bypassing tool restrictions.
  • •Runs `grep` over repository files, which is read-only but still involves shell execution.
How this analysis is made

Examples

Validate a requirement spec
/validate REQ-014
Check the task breakdown
/validate tasks
Gate implementation before merge
/validate implementation

name: validate description: Validate any ADLC phase output before advancing argument-hint: REQ-xxx ID or phase name (spec, architecture, tasks, implementation)

/validate — ADLC Phase Validation

You are validating ADLC artifacts to ensure quality before advancing to the next phase.

Ethos

!sh .adlc/partials/ethos-include.sh 2>/dev/null || sh ~/.claude/skills/partials/ethos-include.sh

Context

  • Active specs: !grep -rl 'status: draft\|status: approved\|status: in-progress' .adlc/specs/*/requirement.md 2>/dev/null | head -20 || echo "No active specs"

Input

Target: $ARGUMENTS

Prerequisites

Before proceeding, verify that .adlc/specs/ exists. If it doesn't, stop and tell the user: "The .adlc/ structure hasn't been initialized. Run /init first."

Instructions

Step 1: Identify What to Validate

  1. If given a REQ ID, locate all artifacts under .adlc/specs/REQ-xxx-*/
  2. If given a phase name, validate the most recently modified artifacts for that phase
  3. Determine the current phase based on what artifacts exist:
    • Spec phase: Only requirement.md exists
    • Architecture phase: architecture.md exists alongside requirement
    • Task phase: tasks/ directory with task files exists
    • Implementation phase: Tasks have status complete or code changes exist

Step 2: Validate Based on Phase

Validating a Requirement Spec

  • [ ] Frontmatter has valid id, title, status, created, updated fields
  • [ ] Description clearly explains what AND why
  • [ ] Acceptance criteria are specific, testable, and use checkbox format
  • [ ] No implementation details in the requirement (belongs in architecture)
  • [ ] Assumptions are explicitly stated
  • [ ] Out of scope items are defined to prevent scope creep
  • [ ] External dependencies are identified
  • [ ] No duplicate or overlapping requirements with existing specs

Validating Architecture

  • [ ] Architecture follows existing patterns from .adlc/context/architecture.md
  • [ ] New ADRs include rationale (not just decisions)
  • [ ] Data model changes are compatible with existing Firestore schema
  • [ ] API endpoint design follows REST conventions from .adlc/context/conventions.md
  • [ ] Service layer follows the layered pattern (routes → services → repositories)
  • [ ] No architectural conflicts with other in-progress requirements

Validating Tasks

  • [ ] Every task has valid frontmatter (id, title, status, parent, created, updated, dependencies)
  • [ ] Tasks form a valid DAG — no circular dependencies (including cross-repo dependencies)
  • [ ] Every acceptance criterion from the requirement is covered by at least one task
  • [ ] Each task lists specific files to create/modify
  • [ ] Tasks are appropriately scoped (not too large, not too granular)
  • [ ] Test requirements are included in task acceptance criteria
  • [ ] Dependencies reference valid task IDs

Cross-repo tasks — only applies if .adlc/config.yml in the primary repo declares more than one entry under repos:. Skip these checks in single-repo mode.

  • [ ] Every task has a repo: field in its frontmatter
  • [ ] Every repo: value matches an id declared in .adlc/config.yml under repos:
  • [ ] No task modifies files outside its declared repo: — all paths under "Files to Create/Modify" live in that repo
  • [ ] At least one task targets the primary repo (even if just spec/doc updates) OR a follow-up confirms the primary needs no code changes
  • [ ] Cross-repo dependencies make sense (e.g., a frontend task depending on a backend task, not the reverse)

Validating Implementation

  • [ ] All task acceptance criteria are met
  • [ ] Tests pass (npm test or equivalent)
  • [ ] Code follows conventions from .adlc/context/conventions.md
  • [ ] No new lint warnings or errors
  • [ ] All requirement acceptance criteria are satisfied
  • [ ] ADLC artifacts are updated (task statuses set to complete)

Step 3: Report Results

  1. Display validation results as a checklist with pass/fail for each item
  2. Categorize issues by severity:
    • Blocker: Must fix before advancing (e.g., missing acceptance criteria, circular deps)
    • Warning: Should fix but won't block (e.g., vague wording, missing edge case)
    • Info: Suggestions for improvement
  3. If all checks pass, confirm the artifact is ready to advance
  4. If blockers exist, list specific fixes needed

Step 4: Recommend Next Action

  • Spec validated → "Ready for /architect"
  • Architecture validated → "Ready for implementation"
  • Tasks validated → "Ready for implementation"
  • Implementation validated → "Ready for /review"

From the file to a first result

The download contains exactly the content displayed on this page. This copy may differ from the latest version in the source repository.

View the source and supporting files

This download contains SKILL.md only. Any scripts, references, templates and dependencies are not included. Get the required files from the source, preserving their directory structure, before continuing.

1. Download the file

Download SKILL.md

From the root of a test project, save the file as .skill-downloads/adlc-phase-validation-ms47vp5z/SKILL.md, or use the command below. This folder is for inspecting the skill before installation.

Command for macOS/Linux or WSL: requires a POSIX shell and curl. On Windows without WSL, use the download button and create the folder manually.

mkdir -p '.skill-downloads/adlc-phase-validation-ms47vp5z' && curl --fail --show-error --location 'https://skills-guides.com/api/skills/adlc-phase-validation-ms47vp5z/download' --output '.skill-downloads/adlc-phase-validation-ms47vp5z/SKILL.md.part' && mv '.skill-downloads/adlc-phase-validation-ms47vp5z/SKILL.md.part' '.skill-downloads/adlc-phase-validation-ms47vp5z/SKILL.md'

2. Check prerequisites

Read SKILL.md before using it. If it references scripts or external services, inspect them and prepare the required access in your test environment.

Original excerpt: Prerequisites

Before proceeding, verify that .adlc/specs/ exists. If it doesn't, stop and tell the user: "The .adlc/ structure hasn't been initialized. Run /init first."

The platforms on this page are catalog indications. The paths below follow tool documentation; they do not prove this skill has been run in those tools. No minimum tool version is confirmed for this listing.

3. Load the skill

After inspection, copy the complete folder into your project using one of the commands below, then open a new tool session. The command stops if a folder with the same name already exists: compare the files before replacing it.

Claude Code
test -f '.skill-downloads/adlc-phase-validation-ms47vp5z/SKILL.md' && test -s '.skill-downloads/adlc-phase-validation-ms47vp5z/SKILL.md' && mkdir -p '.claude/skills' && test ! -e '.claude/skills/validate' && cp -R '.skill-downloads/adlc-phase-validation-ms47vp5z' '.claude/skills/validate'

OpenSkills

OpenSkills must already be installed. The local install/read procedure is checked with OpenSkills 1.5.0, which requires Node.js ≥ 20.6.0. Catalog slugs are not resolved automatically.

Official OpenSkills installation
openskills install './.skill-downloads/adlc-phase-validation-ms47vp5z'

openskills read 'adlc-phase-validation-ms47vp5z'

Reading should display the installed file content and its base directory. It does not run the skill or install its dependencies.

For an agent that reads AGENTS.md, then sync the index. This command changes AGENTS.md: review the diff before using it.

openskills sync

Manual method or another tool: attach SKILL.md to a conversation, or open it in an agent that can read files. Include any required supporting resources. No universal folder is recognized by every tool.

4. Try a first task

Use a short, synthetic input: an API endpoint, a small document or an interface brief. In a new conversation, attach the file or give the agent its path, then adapt this request:

Read .skill-downloads/adlc-phase-validation-ms47vp5z/SKILL.md and apply its instructions to this example: [describe a test input]. Identify any missing files, dependencies or access. Produce the result requested by the skill without inventing data or running unreviewed scripts.

Expected result: output matching the skill instructions, or a precise list of missing prerequisites. A read confirmation alone does not validate a first use. If an external resource is unavailable, that path remains untested.

Related skills